by Michele Laurelli
A framework where two neural networks compete: a generator creates fake data and a discriminator tries to distinguish real from fake.
GANs consist of a generator and discriminator trained adversarially. The generator improves at creating realistic data while the discriminator improves at detection. Used for image generation, style transfer, and data augmentation.
StyleGAN for face generation
CycleGAN for image-to-image translation
Pix2Pix for paired image translation